A cross-platform OCR library based on PaddleOCR v5 and ONNX Runtime that is as small as possible.

Speech-to-text, text-to-speech, speaker diarization, and speech enhancement using Next-gen Kaldi without internet connection [termux android-arm64 build with bundled .so]

High-quality text-to-speech for the web

Node-first OCR interface with support for Tesseract.js, ONNX (PaddleOCR), HappyVertical AI-backed vision OCR, and Unlimited-OCR

Fork of @diffusion-studio/vits-web for easier built-in PiperTTS use.

Compiled Protobuf definitions for ONNX files

Web api for using VITS based models in the browser!

Fast PDF classification and text extraction. Detect text-based vs scanned PDFs, extract text by region with quality checks. Native Rust performance via napi-rs.

A Javascript library for running ONNX models on browsers

Rampart — client-side PII redaction for AI assistants: deterministic recognizers + a 14.7 MB ONNX classifier (transformers.js), default-deny policy, and reversible placeholders. Runs entirely in the browser.

Geti smart tools utils and interfaces

React Native wrapper for sherpa-onnx TTS and STT capabilities

ONNX Runtime detection backend for camera.ui. Runs object detection, face detection and recognition, license plate recognition with OCR, and CLIP semantic embeddings.

MediaPipe hand landmark detection neural network model for Unity Barracuda

ONNX Runtime detection backend for camera.ui with the older CUDA 12 runtime. For NVIDIA GPUs before GTX 1650 (Maxwell, Pascal, Volta) and for systems that stay on CUDA 12. Use the regular ONNX plugin on newer hardware.

ONNX Runtime Plugin for Unity

MediaPipe BlazePalm palm detection neural network model for Unity Barracuda

MediaPipe iris landmark detection neural network model for Unity Barracuda

This module is wrapper over sherpa onnx for tts

MediaPipe face landmark detection neural network model for Unity Barracuda